Doing the project right, and doing the right project
Almost every instrument the industry owns answers the first question. Estimating, cost control, scheduling, gateways, technical assurance, internal audit, all of them examine execution. Whether this should have been the project at all is settled once, early, by fewer people, and becomes progressively harder to ask.
An organisation can be genuinely excellent at delivering projects and have no capability whatsoever for deciding which ones to deliver.
The two are not related, they are not measured together, and they are not usually held by the same people. One of them has an entire industry attached. The other is settled in a handful of meetings, early, by a smaller group, and after the announcement it becomes progressively harder to reopen.
Where the instruments actually point
Consider what a large owner has available to it. Cost estimating and estimate classification. Schedule development and critical path analysis. Earned value. Change control. Risk registers. Technical assurance. Gateway reviews. Commercial and contract audit. Internal audit. Cost consultants. Lenders' advisers.
Every one of those examines execution. Each asks, in its own vocabulary and with real rigour, whether the thing is being done properly.
Not one of them asks whether it is the right thing.
That is not a criticism of any of them, and it is not an accident. An instrument needs something to measure, and execution generates artefacts continuously: reports, valuations, programmes, registers. Selection generates almost nothing except a decision.
Why assurance arrives when the artefacts do
, read: A gate that has never been failed is not a controlWhat is actually required before a shortlist exists
There is a published method for the other question, and it is more demanding than most people who cite it realise.
The Green Book, the UK government's guidance on appraisal, sets out four stages. Rationale and objectives. Generating options and longlist appraisal. Shortlist appraisal. Identifying the preferred option.
Only the last two involve comparing the things most business cases contain. The first stage asks for five elements, and each is a different question:
Case for change, explaining why the organisation needs to act at all. Theory of change, explaining how the proposal will produce the intended outcomes. Business as usual, describing the outcome expected if current arrangements continue and nothing is done. Objectives, which must be specific, measurable, achievable, realistic and time-limited. Strategic fit, explaining how the proposal aligns with the wider objectives of the organisation and of others.
Then, before any detailed analysis, a longlist. The guidance is direct about why it must be wide: a sufficiently broad longlist helps to avoid thinking too narrowly and reduces the risk of overlooking better options.
The rule that does the work
One instruction carries more weight than the rest, and it is the one most often quietly disregarded.
An option must not be taken forward to shortlist appraisal if it does not achieve the proposal's objectives. The exception is the business as usual, which must always be taken forward to shortlist appraisal. It acts as a benchmark against which other options are compared.
Read what that requires. The option of not doing the project has to survive the filter, enter detailed analysis, and be costed and compared alongside the interventions, specifically so there is something to measure them against.
What usually happens instead is that a do-nothing option appears in a paragraph, is described in terms nobody could support, and is disposed of before the analysis starts. That is not a benchmark. It is a formality performed in the direction of a decision already taken.
The 2026 edition also changed the vocabulary in a way worth noticing. Earlier editions spoke of a do minimum. This one specifies business as usual, defined as the outcome expected if current arrangements continue. The difference is not cosmetic: a do minimum invites you to design a deliberately feeble alternative, while business as usual is a factual forecast of what happens anyway, and a factual forecast can be wrong in ways somebody can check.
Theory of change, and the step usually skipped
The second element deserves separate attention because it is the one that most often does not exist in any form.
A theory of change explains the mechanism. Not that the asset will produce the benefit, but how: through what chain of cause and effect, under what conditions, and with what else needing to be true.
Most business cases assert the outcome. Capacity will increase, therefore congestion will fall. Facilities will improve, therefore service will improve. Each of those has at least one hidden step, and the hidden step is usually where the benefit either survives or evaporates.
The practical value of writing it down is that a mechanism can be falsified and an assertion cannot. It also names the conditions the benefit depends on, several of which will turn out to be somebody else's responsibility, which is worth discovering at appraisal rather than in year three.
Why the honest version is hard, and getting harder
The Green Book also says that practitioners should consider whether similar interventions have previously been evaluated, and should use that evaluation evidence where relevant.
That instruction assumes a stock of evaluated outcomes to draw on. The Evaluation Task Force found that two thirds of the UK's major projects have no adequate plan to establish their own outcomes, and that value for money is the least evaluated of the three types.
So the method for choosing the right project depends on evidence that the failure to evaluate the last ones prevents anyone from having.
What the evaluation review actually found
, read: Whether it was worth doing is the least evaluated questionChoosing between programmes, not within one
There is a further gap that the guidance addresses less well, and it is the one that matters most at portfolio scale.
Appraisal as described chooses between options for meeting a stated objective. It is a within-programme instrument. It does not choose between programmes, and on a portfolio the more consequential question is frequently which of several defensible things to do with money that will only stretch to some of them.
Most capital governance approves programmes one at a time, sequentially, each against its own business case. A proposal is compared to its own alternatives and to business as usual. It is rarely compared to the other proposals competing for the same envelope, because they arrive at different committees on different dates.
The consequence is a portfolio assembled by accretion. Every individual approval was defensible. Nobody chose the set.
Reading this from here
The region has spent a decade demonstrating considerable capability at the first question. Programmes of unusual scale have been delivered, and the delivery apparatus that did it is real.
The circumstances have since changed. At a board meeting in December 2024 the Public Investment Fund approved a minimum twenty per cent spending reduction across its portfolio, and several giga-project timelines have been recalibrated. A portfolio being reduced has to choose what to stop, and stopping is the selection question arriving late and in its least comfortable form.
An organisation that has never built the instrument for it will find that the decision gets taken anyway, on the basis of which programmes are most visible, most advanced, or most difficult politically to abandon. Those are real considerations. They are not the same as which ones were worth doing.
The framework for this stage sets out how to examine whether an outcome exists separately from the solution, and whether the criteria were written before the appraisal, which is the difference between a choice and a preference.
Sources. HM Treasury, The Green Book, 2026 edition, for the four stages of appraisal, the five elements of rationale and objectives at paragraph 2.11, the breadth of the longlist at 2.13, critical success factors at 2.14, the treatment of business as usual at 2.16, and the instruction on using prior evaluation evidence at 2.9. Evaluation Task Force, Government Major Projects Evaluation Review, April 2025, for the evaluation coverage figures. HM Treasury, Supplementary Green Book Guidance, Optimism Bias. ISO 21502:2020 for benefit management as a project practice. The Public Investment Fund spending reduction was reported by Arabian Gulf Business Insight in March 2025, describing a board decision taken in December 2024 and attributed to unnamed sources rather than to a statement by the fund.
Related reading
The largest contributor to optimism bias is the document that starts the project
HM Treasury publishes uplifts by project type, from 24 per cent on a standard building to 200 on equipment and software. The instruction is to start at the upper bound and come down only against verified evidence. The single largest contributory factor it names is not the contractor, the ground or the weather.
ReadWhether it was worth doing is the least evaluated question
The UK government reviewed evaluation across the 227 largest projects it runs, together worth £834 billion. A third have a plan capable of finding out whether they worked. Of the three kinds of evaluation, the one asking whether the benefits justified the cost is the weakest, and the reason is that it has to be designed before the project starts.
ReadA gate that has never been failed is not a control
Assurance before anything is built costs less and changes more than assurance anywhere else in the lifecycle. It is also the assurance almost nobody buys, and the reason is not that owners are careless. It is that inception produces very little of the sort of evidence an assurance function knows how to examine.
Read